Executive Summary
Onetree partnered with ConnectAmericas (Inter-American Development Bank) to design and develop ADA, an AI-powered business matchmaking platform that transforms how companies connect, collaborate, and create opportunities across Latin America and the Caribbean.
As ConnectAmericas’ ecosystem grew to hundreds of thousands of companies, manual partner discovery and static search tools became a bottleneck, limiting the quality, relevance, and scalability of business connections. The challenge was not only to improve search accuracy, but to enable meaningful, intent-driven matches at scale, beyond one-to-one meetings and in-person events.
ADA was conceived as a conversational, human-centered AI assistant, designed to understand real business goals through natural language interactions rather than rigid forms or filters. Powered by generative AI, semantic search, and affinity scoring, the platform delivers highly relevant, explainable matchmaking results tailored to each company’s context.
Built as an MVP in approximately three months, ADA combines robust AI architecture, structured data preparation, and a continuous feedback loop, enabling ConnectAmericas to scale high-quality business connections while positioning the platform as a regional leader in AI-driven matchmaking for trade and investment.
The Challenge: Scaling Relevant Business Connections
As ConnectAmericas continued to grow as a regional business ecosystem, the platform faced a critical challenge: scaling meaningful business connections without sacrificing relevance or quality.
With hundreds of thousands of registered companies across Latin America and the Caribbean, partner discovery relied heavily on manual searches, static filters, and keyword-based matching. While functional at smaller scale, this approach made it difficult for companies to identify truly relevant partners, often resulting in missed opportunities and low conversion from connection to collaboration.
In addition, ConnectAmericas’ matchmaking model was largely tied to one-to-one meetings and in-person events, limiting its reach and impact. The organization needed a way to extend high-quality matchmaking beyond physical events, enabling always-available, virtual business connections that could scale across regions, industries, and languages.
The core challenge was not just improving search, but reimagining business matchmaking as an intelligent, personalized, and scalable experience, capable of understanding business intent, adapting to diverse contexts, and supporting the platform’s long-term mission of facilitating trade and investment at scale.
The Solution: Introducing ADA, an AI-Powered Matchmaking Platform
To address the challenge of scaling meaningful business connections, Onetree partnered with ConnectAmericas, an initiative of the Inter-American Development Bank, to design and build ADA, an AI-powered matchmaking platform purpose-built for trade and investment ecosystems.
ADA was designed as a conversational, intent-driven experience, shifting partner discovery away from static filters and keyword searches toward natural language interactions that capture real business goals, priorities, and constraints. This approach enables more accurate and relevant matches while reducing friction for users navigating complex business ecosystems.
The platform was developed as a minimum viable product (MVP) in approximately three months, in close collaboration with the Microsoft AI Lab, one of only five global Microsoft AI Labs dedicated to accelerating applied AI innovation. This partnership supported AI architecture design, model selection, and an in-person sprint that led to the first fully functional version of ADA.
By combining a human-centered product experience with an intelligent matchmaking approach, ADA enables ConnectAmericas to scale high-quality business connections across regions, industries, and languages, while continuously improving through feedback and real-world usage.
While ADA leverages foundational large language models (LLMs), its real differentiator lies in the structured preparation of business data, intent-aware prompt design, carefully curated context, and a continuous feedback loop. This approach ensures that the platform delivers reliable, explainable, and scalable matchmaking results, positioning ADA as a purpose-built AI product rather than a generic or out-of-the-box AI integration.
Key Capabilities
- AI-powered business matchmaking based on real business intent, not just keywords or static filters.
- Conversational interface powered by generative AI that guides companies through natural language interactions to uncover goals, priorities, and opportunities.
- Personalized matchmaking algorithms that consider sector, location, company size, and specific business needs.
- Affinity scoring and match prioritization, helping users focus on the most relevant and high-potential connections.
- Multiple matchmaking modes tailored to different use cases, including suppliers, clients, and strategic alliances.
- Multilingual experience, supporting Spanish, English, Portuguese, and French to enable cross-regional connections.
- Continuous feedback mechanisms, allowing users to rate recommendations and helping the system improve over time.
- Analytics and insights that support better networking decisions and help optimize matchmaking strategies at scale.
Technical Foundations
- Semantic data layer, built on vectorized company profiles to support intent-based reasoning and AI-driven retrieval.
- AI prompt and context orchestration, designed to guide matchmaking logic and ensure consistent, explainable responses across use cases.
- Data preparation and enrichment pipelines, transforming heterogeneous structured and unstructured business data into AI-ready inputs.
- Context-aware matchmaking models, designed to support distinct partner discovery scenarios such as suppliers, clients, and strategic alliances.
- Match evaluation models, combining multiple signals to assess relevance, confidence, and strategic alignment between companies.
- Explainability layer, generating contextual justifications that support transparency and trust in AI-assisted recommendations.
- Multilingual AI configuration, enabling consistent reasoning and responses across Spanish, English, Portuguese, and French.
- Headless backend architecture powered by Directus, providing a flexible and decoupled foundation for managing business data and platform content.
- Scalable AI architecture, designed and validated in collaboration with the Microsoft AI Lab, supporting rapid experimentation, iteration, and long-term growth.
Results & Impact: Smarter Business Connections at Scale
Thanks to ADA, ConnectAmericas has transformed how businesses interact on the platform. The new system delivered measurable improvements across the ecosystem by enabling more relevant, efficient, and scalable business matchmaking.
- Significantly increased the quality and success rate of business connections, enabling more relevant and meaningful matches.
- Reduced the time required to identify relevant business partners, streamlining partner discovery at scale.
- Enhanced the user experience through a dynamic, intuitive, and value-driven conversational interface.
- Positioned ConnectAmericas as a pioneer in AI-powered business matchmaking across Latin America and the Caribbean.
This project highlights Onetree’s commitment to building innovative, scalable, and human-centered technology solutions that address complex challenges with real-world impact.